Artificial Intelligence can be a very useful tool to help you with the burden off the shoulder when it comes to the mundane and rival tasks. Some of the companies like the Google have also been used the machine learning and AI to even emulate people for some of the simple tasks such as sales appointment, book a table in a restaurant with the products such as the Google Duplex.
AI algorithms have also been used to animate the 2D objects such as the talking with the heads of people, but sometimes it also requires a huge set of a database of samples to train the neural network to adapt to the objects or in this case, the personal manners and behavior.
This process even just got a lot much simpler thanks to the Samsung AI center Moscow Chapter, which as of now published reveals that a highly realistic talking head model which can be obtained by feeding the algorithms with as little as a single sample photo.
The research paper which has been titled as the ‘Few-Shot Adversarial Learning of Realistic Neural Talking Head Models’, from the Skolkovo Institute of the Science and Technology in Moscow, founds its way to subreddit along with the demo videos which even shows the Artificial Intelligence in action. The algorithm uses techniques of using landmarks which is a skeletal framework of a person’s face learned by crawling with the help of a large dataset of videos, that is being used as the backbone for the synthesized face.
Such a technique can be used to bring with some of the more realistic looking ‘Memojis’ which an Apple version of the Animoji are but of yourself. The creative potential of this is even endless. Moreover, such technology can also be misused if fallen in any of the wrong hands.
Discussion about this post